Meta expands Instagram teen safety features with nudity protection in DMs

Meta has announced expanded safety protections for teenagers on Instagram, including a global rollout of technology that automatically blurs nudity in direct messages and warns users before they send or open such images. The nudity protection feature, which uses on-device machine learning to analyse image content without Meta seeing the photos, will be enabled by default for users under 18 worldwide. The company is also implementing account restrictions that prevent new accounts and those exhibiting suspicious behaviour from seeing teens in follower lists or being suggested teen accounts, and has introduced video-selfie verification to confirm user age when suspicious activity is detected.

Meta's on-device nudity filter addresses the moment an image is sent or opened, but sextortion and grooming campaigns build over multiple messages through identifiable patterns of escalation, flattery, isolation and coercion long before explicit content appears. Guardii's anti-sextortion and anti-grooming detection modules, developed by a Meta Business Partner, analyse the shape and rhythm of a conversation in real time across platforms including Instagram, Snapchat and Discord, identifying sustained targeting and the shift from benign chat to manipulation without requiring access to message text or images. That approach intercepts the predatory relationship at the approach and escalation phase, surfacing the threat to a parent or safeguarding professional while the child can still be protected, rather than intervening only after explicit material has already been exchanged.